1. INTRODUCTION

In the summer of 1999 a team of nine students1 andthe present writers measured and inventoriedtravertine building elements and non-architecturalobjects in Ostia, its cemeteries outside the PortaRomana and the Porta Laurentina, and in IsolaSacra. With the use of dBase III+, more than 3000objects were marked on maps and registered, includ-ing the dates of buildings as proposed by G. Calzaand others in Scavi di Ostia I (1953), P. Pensabene(1972), Thea Heres (1982), C. Pavolini (1983) andI. Baldassare (1996).Travertine was used for stairs, thresholds (more than1200), doorposts, lintels, door- and windowframes,foundation stones, substructures, stereobate, tombwalls, corner and pillar bases, buttresses,2 bases,columns, capitals (Pensabene 1972), capital-cush-ions (It. cunei/pulvini), pillars, buffer stones (on thecorners of buildings), corbels (consoles), springers,corner- (on the frontal corners of doorways or pil-lars) and bar stones (centralized at the inside ofdoorways or pillars to fix doorbars), keystones,floors, latrine seats, benches, gutters, small sluicepillars in drainage-channels, intramural blocks,pavings, sidewalks, curbstones, and centralizedblocks in basalt streets.Travertine was also used in non-architectural ele-ments such as cippi (e.g. boundary stones), inscribed(usually funerary) panels, well mouths, putealia,small altars, socles for statues, reliefs, poundingblocks, weights,3 and rectangular containers for ashurns.

2. WRITTEN SOURCES

Vitruvius defined travertine (Lat. Tiburtinus lapis)as a middle-hard stone from the region of Tibur(modern Tivoli), able to withstand damage fromheavy loads and bad weather, but susceptible to firedamage because of its dry and porous character (DeArch. 2, 7, 1-2):Tiburtina (sc. saxa) vero et quae eodem genere suntomnia, sufferunt et ab oneribus et a tempestatibusiniurias, sed ab igni non possunt esse tuta.....Pliny (N.H. 36, 48, 167) gives similar information:

Tiburtini (sc. lapides), ad reliqua fortes, vapore dis-siliunt.(Travertine is split by heat, though it stands up tothe other forces). He adds that when Cicero saw themarble walls of the Chians, which were meant as ashow piece for their visitors, he remarked: ‘I shouldbe much more amazed if you had made it of stonefrom Tibur.’ According to Strabo (Geogr. 5, 3, 11)travertine was easily transported by the Anio, a nav-igable river which flows into the Tiber. The refer-ences to travertine in the literary sources are veryscarce. It is clear that Vitruvius and Pliny mentiontravertine because of its load bearing ability.Nowhere, however, do they discuss its use in build-ings or any of its other functions. As we shall see,travertine was not only used for support.

3. QUARRIES

Roman age travertine quarries are located west ofTivoli 2.5 km from Tenuta del Barco, betweenTenuta Martellone and the Tivoli Mountains.4 Inantiquity travertine blocks were transported by car-riage over twenty-two kilometers to Rome on a road6.5 m wide.Travertine is a calcareous sedimentary rock. Hotsprings deposited it in layers on the ground duringthe Quaternary period of the Middle Pleistocene

169

1 Rebecca van de Berg (bases), Vincent Deurwaarder (corner-stones), Bart Corver (centralized streetslabs), Guiot C.Dûermeijer (thresholds with relief), Carl J.A. van Hees (columns),Giseke R.M. Hopstaken (thresholds without relief), Eric Norde(corbels), Gerdine M. de Rooij (staircases), Marloes P.H. vander Sommen (sidewalks) and some volunteers. Preliminarybibliographical research on travertine in ancient Rome was doneby David Murray (Bowdoin College, USA). He and drs GwenTolud corrected the English text. Many thanks to all, and espe-cially to Mrs. Dr A. Gallina Zevi, Soprintendente di OstiaAntica, who gave permission to the research and offered unfor-gettable hospitality. Thanks also to dottori Bedello, Belfiore,Germoni, Izzi, Marinucci, Miraglia, Panariti, Pellegrino, Valloc-chia (Sopr. Ostia Antica). The research did not include PalazzoImperiale which will be published by dr Joanne Spurza (NewYork; see ArchLaz 10 (1990) 157-163). All photographs havebeen made by the authors.2 Rickman 1971, 60, pl. 32.3 CIL XIV, suppl. I, 5316, 1, 4.4 Mari 1984, 359-370.

BaBesch 75 (2000)

Tiburtinus lapisThe use of travertine in Ostia

L.B. van der Meer and N.L.C. Stevens

Latial volcanism. The formation process involves acristallization of calcium carbonate (CaCO3) fromwater containing calcium bicarbonate (CA(HCO3)2)while carbon dioxide (CO2) is released. This is aprocess comparable to the growth of stalactites andstalagmites. When first quarried travertine is soft,but as it drys it becomes more sturdy. The layersmentioned above vary in height from 45 to 60 cm. InRoman times cubic blocks with sides ranging from120 to 180 cm were sawn from the rock. Accordingto Lanciani the Romans would have transported c.5.5 million tons of travertine to Rome. For theAmphitheatrum Flavium alone c. 1000 tons weretransported daily by c. 150 carriages during a periodof four years.

4. TRAVERTINE IN OSTIA

The use of travertine in Ostia raises many questions:1. When was it first used?2. How long was it used?3. When was it popular and why?4. How was it used and in which context? What are

the reasons for its use?

5. Is the use of travertine functional, decorative orboth? Has it been a status-symbol?

6. Did it replace the more expensive marble?7. Was there any standardization of building ele-

ments?

There is no evidence for the use of travertine inOstia’s earliest settlement, the so-called Castrum,which was recently dated to the end of the fourth orthe beginning of the third century B.C. Travertinealso fails to appear in the oldest known domus atOstia, built after the second Punic War and beforethe period of Sulla.5The oldest non-architectural objects may be sixboundary stones (cippi), standing in one line alongthe northern side of the eastern Decumanus, betweenVia dei Molini and a place just to the east of thePorta Romana, one of the three main entrances of theso-called Sullan wall.6 They stand c. 600 m (c. 2000Roman feet) apart from one another at irregularintervals, more than one metre below the level ofthe street, which was raised in the period of

170

5 See A. Martin, in: Gallina Zevi/Claridge (1996) 19-38.6 Meiggs 1973, 32, 472, 594. NSc 1910, 233.

Fig. 1. Two boundary cippi, Decumanus Maximus (II.9.2)

Domitian. They show identical inscriptions (fig. 1,to the right): C.CANINIVS.C.F.PR. VRBDE.SEN.SENTPOPLIC.IOVDICwhich means that Gaius Caninius, son of Gaius,pr(aetor) urb(anus), based on a decision of theSenate, ordained that the area between the cippi andthe Tiber was public, this means forbidden for pri-vate building (CIL XIV 4702). This (ager) poplicos(publicus) was later restricted by another travertineboundary stone, put directly beside the 5th Caniniusstone (on the east), bearing the inscription (fig. 1, tothe left): PRIVATVMAD TIBERIMVSQUE ADAQUAMwhich means: private area to the Tiber just to the wa-ter (CIL 4703). The Caninius cippi date from c. 150to c. 80 B.C. Scholars disagree on a more specific datefor the cippi, but travertine was first used at Rome in109 B.C. in the Mulvian bridge.7 So the date of thecippi may be restricted between c. 109 and 80 B.C.

The northern bank of the Tiber, now ‘fiume morto’,had at least five travertine boundary stones, placedby C. Antistius Vetus and other curatores riparumet alvei Tiberis (surveyors over the banks and bed-ding of the Tiber). G. Barbieri dates them after 23A.D., the year in which Antistius became consul.8Other early boundary stones have been found in theSanctuary of the Four Republican Temples in reg. II(77 x 24 cm). The four cippi bear the inscriptionI.O.M.S. (Iovi Optimo Maximo sacrum) (CIL XIV4292). They have been dated to the period betweenCaesar and Claudius.9Another two exceptional cippi have been found tothe west of the horrea (V, i, 2) which date to about50 A.D. Both have the same inscription: SEMITAHOR P R I (cross-road/path of the horrea; the mean-ing of the unique abbreviation P R I is unknown).10

171

7 J.B. Ward Perkins, EAA IV (1961) 864-5. About the privatum-cippus, see Meiggs 1973, 472.8 Meiggs 1973, 115, 594. SO I, 62.9 Meiggs 1973, 346.10 Bakker 1994, 197-8; Bakker (ed.) 1999, 113. Bakker suggests:‘principium regionis I’, or ‘primae regionis initium.’ CIL XIV,352 confirms that Ostia had at least 5 regiones. R.E.A Palmer

Fig. 2. Temple of the Round Altar, interior (I.xv.6)

Thus they mention the only name we know of anOstian by-way, which runs north-south from theeastern Decumanus meeting the southern CardoMaximus not far from the Porta Laurentina.As for the funerary use of cippi, the tombs in thecemetery outside Porta Laurentina offer the mostinteresting examples. Standing in front of tombs,they are rounded on top and they mention the namesof the deceased. The cemetery can be dated roughlyfrom the end of the Republican period to the end ofthe first century A.D. As an alternative, inscribedpanels inserted into the facade, have also been used.Travertine panels frequently occur in the cemeteryoutside Porta Romana, usually mentioning the nameof the deceased and the measures of the front andsides (in Roman pedes) of the tomb.11 Most tombs canbe dated between the period of Augustus and c. 100A.D. Marble panels are rare here. A socle of travertineinscribed with the words Herculi HermogenianoSacrum and dating to about 50 A.D. was found at ashort distance from the Tomb of Hermogenes.12

It should be noted that in Isola Sacra, which was inuse mainly during the second century A.D., traver-tine cippi are completely absent and only marbleinscriptions occur.This implies that c. 100 A.D. there must have beena shift from travertine to marble panels. The oldest travertine building elements are visiblein the monumental Temple of Hercules (I.xv.5),which R. Meiggs dates to the closing years of thesecond century B.C.13 The temple’s terminus antequem is c. 70 B.C., which is the latest possible datefor the famous haruspex-relief, one of the marbleex-votos found near the temple. The two steps of thecrepidoma on the sides and on the back, and theeight wide steps of the stairs on the front are oftravertine. The rest of the exterior of the podium hasbeen built in opus quasi-reticulatum. To the south-west of the Hercules temple, in the triangular area,stands the so-called Temple of the Round Altar,which also may date to the first century. On a lowerlevel, just in front of the temple and partially underit, three reused blocks of travertine flank a tufa altar(now replaced by modern copies). They bear Greekinscriptions14 and were originally used as statuesocles. Zevi dates them to the period of Sulla.15 Inorder to be transported from Greece to Ostia, thethree statues were removed from their originalsocles and were replaced with new travertine soclesinscribed with their original texts once they reachedOstia. In the Augustan period, the threshold anddoorposts at the entrance of the temple wererestored to marble. However, for the substructure ofthe threshold, travertine was used. The threshold’sends are distinctly raised (fig. 2) like separate sideblocks, a phenomenon occuring frequently in latercenturies (fig. 3). Even in the fourth century A.D.travertine side-blocks flanked marble thresholds.Other Republican temples do not show travertineelements probably because they underwent modi-fications during the Imperial period.Of particular interest are the remains of a Sullandomus behind the (later) Sede degli Augustali. Theperistyle has slender Tuscan columns of tufa, but thecolumns on the four corners are travertine.16 A similarphenomenon of aesthetic corner accentuation is

172

(JRA 1996) suggests: pri(vata). F. Coarelli suggests: ‘populiRomani iussu.’11 Funerary inscriptions on travertine cippi or panels from Ostiaare mentioned in CIL XIV, 707, 717, 718, 978, 1094, 1120,1301, 1387, 1477, 1477a, 1685, 1757, 1871; CIL XIV, Suppl. I,4287, 4494, 4778 (?), 4874, 4881, 4882, 4893, 4919, 4921,4938, 5034, 5099, 5170, 5182, 5228.12 Meiggs 1973, 350. CIL XIV, 4287.13 Meiggs 1973, 581-2. Zevi 1976, 54.14 W. Helbig, Führer IV (1972), nr. 3388.15 Zevi 1976, 60.16 SO I, 111.

Fig. 3. Detail of threshold (IV.i.9)

visible in the Claudian Horrea of Hortensius(fig. 4).17 It was evidently a tradition which lastedmore than a century.It should be noted that the Temple of Roma andAugustus, one of the first temples built in theImperial period under Tiberius, incorporated opusreticulatum and marble, but no travertine. It is evi-dent that later temples use relatively little travertine.The Capitolium, built during the Hadrianic period,has split windows of travertine only. The theater builtby Agrippa in 12 B.C. shows a strikingly ambiguoususe of marble and travertine. Enlarged in about 196A.D. the outside of the theater’s cavea has traver-tine pillar bases and stairs but the seats inside aremarble. This is an example of selective and hierar-chic use of building materials. Marble accentuatesthe interior while the use of travertine on the exte-rior is more economical and limits the risk of dam-age from the street. This type of design may occuras early as the Augustan period. Supporting thisargument are two tombs outside Porta Marinawhich show a similar combination of materials. Thefamous monumental Mausoleum of C. CartiliusPoplicola, which dates between c. 30 and 25 B.C.,has travertine sides, a tufa back, and a marble facade

with inscriptions and reliefs.18 On one hand traver-tine works as a substitute for more expensive mar-ble, while on the other hand it was considered moreprestigious than tufa. The other contemporaneousmausoleum, just outside the Porta Marina andunfortunately without an inscription, is almostentirely travertine. Only the lost tholos was of Lunamarble. Here too the most eye catching part wasaccentuated by the most expensive material.Up until the age of Augustus porticoes were madeusing tufa. Although the Emperor found Rome as acity of bricks and left it as a city of marble, the samecannot be said of Ostia. Nevertheless some indica-tions of face lifting the streets can be seen at thistime. At least three porticoes with columns of trav-ertine were built: 1) just in front of the (later) Termedel Faro along the southern Cardo Maximus (IV.ii.1;SO I, 108), 2) on the corner of the western decumanusand the Via degli Aurighi (III.iii.1; SO I, 109) and3) at the western side of the Via Epagathiana(I.xiv.9; SO I, 111) in front of the later Terme diButicoso. A facade with Tuscan columns was built

173

17 Meiggs 1973, 45.18 Zevi 1976, 56-9.

Fig. 4. Horrea of Hortensius (V.xii.1)

174

Fig

. 5. P

lan

of d

istr

ibut

ion

of s

tree

tsla

bs (

from

Pav

olin

i 19

86 w

ith

addi

tion

s by

B. C

orve

r)

in the first half of the first century A.D. in front ofa row of shops (I.x.2) along the eastern side of theVia Pomeriale facing the Macellum.19 A second por-tico along the eastern side of the southern CardoMaximus, which was incorporated into the laterTrajanic horrea may date to c. 50 A.D. (I. xiii.4).20

The following architectural elements will be dis-cussed separately: centralized streetslabs (coverstones), sidewalks, corner and pillar bases, thresh-olds with and without relief, stairs, corner and cen-tralized bar stones, springers and corbels. Completedoor-frames, especially those in the tombs of theIsola Sacra cemetery will be discussed more fullybecause of their interesting proportions.

4.1 Centralized streetslabs

Streets in Ostia, both from the Republican andImperial periods, have been made of basalt stones.The streets of the higher, present day level date tothe time of Domitian, about 100 A.D. In the middleof the streets about fifty rectangular travertine slabs,with an average size of about 83 x 42 cm, can beseen; usually two to four slabs lie side by side.Marble slabs can also be seen in isolated incidencesor incorporated into the travertine paving. They maydate to the third century or later when marble wascheaper. The largest concentration of marble slabsis in and near the city center, most frequently in theeastern part of the Decumanus Maximus (fig. 5).Because some slabs have openings and are lyingabove holes, it seems likely that they cover drainshafts. An isolated slab next to the Baths of Mithras(I.xvii.2) and a cluster of drains situated in the Via dellaForica near the public toilet at the rear of the Caseg-giato dei Triclini have a similar function. Some slabsare situated at a crossing and therefore have a strat-egic function in the drain control.The following is a count of travertine streetslabs andtheir specific locations: Decumanus Maximus: 10,Via degli Aurighi: 1, Via delle Corporazioni: 2, Via diDiana: 1, Via della Foce: 5, Via delle Volte Dipinte: 1,Via della Fontana: 2, Via dei Molini: 5, Via delleTerme del Mitra: 1, Via Epagathiana: 5, Via delPomerio 4, Via dei Cippi 5, of which 4 are side byside, and Via della Forica: 8 (in an irregular cluster).Therefore, streetslabs are most commonly found inand near the city center, and especially at crossings.While they are frequently found in the eastern partof the Decumanus Maximus, they are absent in theCardo Maximus. Possibly the latter street had nodrain. The lengths of the blocks vary from 41 to 118cm and the widths from 11 to 87 cm. There are nofixed relations between lengths and widths. Theratio of their measures vary from 5:1 to almost 1:1.The measures have no relation to the Roman foot.

Because there is no trace of standardization, the coverstones must have been made ad hoc. Aside from therectangular examples small round travertine stoneshave been used to repair broken basalt stones in theVia del Tempio Rotondo, Via di Diana and Via deiCorporazioni. In conclusion, rectangular travertinestreetslabs were used because of their white color tocover and mark drain shafts. They could evidentlysupport the weight of carts.

4.2 Sidewalks

Unfortunately sidewalks of travertine, complete orpartial, have been preserved in a rather fragmentarystate. If several materials have been used, the curbmay be travertine, while the rest is tufa, brick, tilesor basalt. Sometimes sidewalks begin with traver-tine and end with basalt blocks and they may havea foundation of brick.It is difficult to see whether they correspond to abuilding or part of a building as is often the case inPompeii as Catherine Saliou has recently demon-strated.21 Sometimes this is the case (see hereafter,survey nrs. 9, 10c); but this is not necessarily true,especially when the sidewalk is longer or shorterthan a parcel. If the sidewalks belong to a buildingthey would have the following functions: a) the pro-tection of building, b) ornamental, possibly relatedto status.If they are part of the urban lay-out they would havethe following functions:a) separation of pedestrians from mobile traffic, andb) restriction of the number of vehicles. Remains of

sidewalks, that is to say raised paved areas along-side the facades of buildings, are visible in thefollowing streets:

1. Decumanus Maximus, along the Caseggiato deiTriclini (I.xii.1).

2. Ibidem, near Porta Marina, along the Domus delNinfeo (III.vi.2) and along the Inn of AlexanderHelix (IV.vii.4).

3. Ibidem, along the Portico degli Archi Trionfali(V.xi.7) and along the Theatre.

4. Via degli Aurighi; by the corner of the Via delleVolte Dipinte (III.ix.23).

5. Via delle Corporazioni, along the Insula dell’Ercolebambino and the Insula del Soffitto dipinto(II.vi.4; II.vi.6 (fig. 6) and the theater; originallythis street probably had complete sidewalks onboth sides. A series of high travertine blocks, pos-sibly of a pavement, remains at the crossing withthe Via della Fullonica.

6. Via di Diana (I.ii.6).

175

19 SO I, 111.20 SO I, 125. Meiggs 1973, 45.21 Saliou 1999, 161-218

176

Fig. 6. Sidewalk (II.vi.4)

Fig. 7 Via Epagathiana. Left side: bases and bar stones (I.xiv.8-7)

7. Via delle Volte Dipinte (III.ix.22).8. Via della Fortuna Annonaria (II.vi.7).9. Via delle Terme del Mitra (I.ii.17).10. Cardo Maximus (IV.iii.1; V.i.1 and I.vi.1).11. Via dei Balconi (I.iv.1).12. Via della Caupona (IV.ii.2).13. Via del Tempio Rotondo (I.xi.1 and IV.iv.7).From this survey it becomes clear that most side-walks occur in Regio I. The Decumanus Maximus hasthe largest number of curbs. This can be explained bythe heavy traffic that used this thoroughfare. Becausethe streetlevel was heightened under Domitian, theterminus post quem for most sidewalks is c. 100 A.D.There are no traces of standardization. The widthvaries from 3.75 to 5.7 pedes and the height variesfrom 0.3 to 1.5 pedes. If only the curb is travertine,its width is about 1 pes, like the small stairs whichsometimes interrupt the curb. The longest remain-ing travertine sidewalk measures c. 7.7 m.

4.3 Corner and pillar bases

Eight premises have bases at one or more corners ofthe facade of one or more buildings. Most examplesare visible in Regio I, along large, busy streets anddate to the period of the big building boom, the timeof Hadrian. These bases may have had three func-tions. First, if there are two corner bases in line, theymay have been intended to indicate the beginningand end of the facade of a building. Thus, they ear-marked the front side of a parcel and its building(s).They tend to project at the front and slightly at thesides and sometimes at the rear. In this case, the sec-ond function may have been to protect the buildingagainst the damage of passing traffic. And finallythey supported the corner walls. Since not everybuilding from the Hadrianic period has corner bases,status may have also played a role. In some casesthe corner bases flank a line of pillar bases in a por-ticus.

Survey.1. The oldest corner base can be found on the leftcorner of building I.xiv.8 (near the Terme diButicoso) on the western side of the ViaEpagathiana. In line with it are pillar bases in frontof I.xiv.8 without panels and I.xiv.7 with cut outpanels (fig. 7). At the end of the porticus, there mayhave been a now lost second corner base to mark theparcel of two buildings. The pillar bases have verti-cal grooves, which may have been for drainage.The following five places show items from theHadrianic period (ca. 120 A.D.).2. Three items, which are in-line, are visible on theleft corner of the rear side of the Caseggiato delTermopolium (I.ii.5), the right corner of the so-called

Basilica (I.ii.3) and the right corner of I.ii.1 (over-built by the Ninfeo), and all along the northern partof the Decumanus Maximus in the center. 3. Two items in-line are on the left and right cornersof the Caseggiato del Larario (I.ix.3).22

4. Two small items are in-line on the southern sideof a building to the west of the Domus di Amore ePsiche (I.xiv.5).5. Three items are in-line on the western side of theVia della Foce on left and right corners of III.xvi.6and on the left corner of III.xvii.5. The latter twoflank the opening of the Via del Serapide.6. Probably from the Hadrianic period are two cor-ner and two pillar bases in a lost building, V.xiii.1,along the southern side of the Decumanus Maximus.7. Dating to the Antonine period are the cornerbases on the ends of a curved porticus with pillarbases along the Cardo Maximus, on the northerncorner of I.xii.7 (two blocks) and on the northerncorner of I.xii.14 (three unequal blocks) and allalong the south-western side of the Terme del Foro.The pillar bases have vertical grooves.8. The corner bases in-line with the left (two blocks)and the right corner (three blocks) of the longersouthern side of the Tempio Collegiale (I.x.4) dateto the end of the Severan period.The corner bases are 74 to 180 cm long, 63 to 146cm wide, and 32 to 52 cm high. The pillar bases are60 to 197 cm long, 56 to 181 cm wide, and 13 to120 cm high. In both cases there are no fixed rela-tions between the three measures, which may be dueto differences in the level of terrain, the size of avai-lable blocks, and the building speed. There are notraces of standardization. Both types of bases havea liminal function in common. The distancesbetween corner bases vary from 1420 to 6377 cm.The distances between pillar bases range from 300to 700 cm, dependant on corridors and shops orother rooms.The phenomenon of corner and pillar bases disap-pears after c. 220/230 A.D.

4.4 Thresholds without relief

Flat thresholds, without edges, grooves or pivotholes, can be found chiefly in horrea, guild buildings,and (larger) dwelling complexes. Flat thresholds didnot support doors. They mark transitional placeswhere people could move unhindered from one spaceto another. They are level with one or two floors. Trav-ertine thresholds seldom appear in the same build-ing with marble ones. It is evident that the architectschose between travertine and marble. Unprocessed

177

22 L. Socca, Nuovi elementi per l’interpretazione del Caseggiatodel Larario a Ostia, AC 46 (1994) 421-440.

thresholds are often found with travertine thresholdswhich have reliefs. Flat thresholds are most fre-quently found in the Regions I, III and IV. Generallythey can be dated to the second century A.D.Flat thresholds are found in the following contexts:1. the official main entrance of a main building;2. the entrance of a lane or corridor which divides

two complexes;3. between two rooms indoors;4. on the border of a complex as delineation;5. the edge of a porticus on the border of a building

and a street;6. the entrance of a vestibule giving access to a stair-

case leading to the first floor; or7. the entrance of a room without another entrance.The measures of flat thresholds were not standard-ized. Their measures depended mostly on that of theentrance. The flat threshold functioned liminallymarking the entrance.23

4.5 Thresholds with relief

Processed thresholds are the most common type.They all have a raised edge (4 to 10 cm thick) at thefront and one or two pivot holes in the corners oftendirectly behind the edge.Sometimes there are raised edges on the smallersides or the thresholds may be framed by raisedtravertine blocks. The most frequent type, whichconsists of several slabs, occurs in the facades oftabernae (shops, workshops, bars and so on). Usually,there is a round pivot hole on the right where thecardo of the door turned. Behind the raised edge isa groove into which the boards of a shutter could bepushed.24 The groove ends where the space of thedoor begins. The slab, where the door could be turned,has a cut-away level. In this manner the boardscould be inserted into the groove.25 The boards wereheld together by a horizontal bar which was fixedinto a bar-hole in the left wall of the entrance.26

There are three other types of thresholds:1) small thresholds with one pivot hole, usually onlyone door, occuring in small rooms without otherdoors like in bedrooms;2) larger thresholds with two pivot holes, for twodoors, have been used at main entrances of build-ings and corridors;3) the largest thresholds have two pivot holes andtwo square holes in the centre for vertical bars. Thistype occurs in buildings where safety was veryimportant, for example in the main entrances of theCaserma dei Vigili (II.v.1) and the Caseggiato delLarario (I.ix.3).Finally, horrea present another type of threshold. Inthe Horrea of Hortensius (V.xii.1; c. 40-50 A.D.) theentrances to the storerooms show blocks with apivot hole on both sides. In the centre a connecting

piece of travertine is absent. Rickman and Meiggspresume that bricks, which are now missing, wereused for the central part.27 It seems more probablethat the central space under the doors was left openfor fresh air. In the Grandi Horrea (II.ix.7) thethresholds are solid, with a marginal check, twopivot holes, and one or two bar holes.28 These thre-sholds, framed by raised side-blocks, may have beeninfluenced by an older type.An exceptional threshold occurs in the Caseggiatodei Molini, in the passage between rooms 7 and 19(fig. 8). The side slabs have a pivot and bar hole,while the central part is basalt. This may have beendone to withstand the exceptional pressure of don-keys, loads of grain, etc.29

In general, there are no fixed standards for the lengthof thresholds. These depend on the width of the door-way. The width, however, may correspond with thethickness of a wall which is often two pedes (c. 60 cm).

4.6 Stairs

The oldest and most monumental stairs in Ostia arethose on the front of the Hercules temple (c. 100 B.C.)mentioned above. The smallest stairs can be foundin the curbs of the sidewalks, which were also men-tioned above. The buildings of the second centuryA.D. have small external stairs corresponding withlarger stairs which lead to the first floor. The smallsteps, varying in length from 60 to 130 cm, bridge thegap between the street and groundfloor-level. One ofthe latest external stairs can be found in front of theentrance of the Domus of Amore e Psiche (c. 325-350A.D.). Inside this domus only small capital cushionsare travertine, while marble dominates the floors andwalls. As for the internal stairs in second centuryhouses, some interesting features of standardizationcan be noted. The width or depth of most treadsvaries from 27 to 33 cm, while the measures of 28,29 and 30 cm, about one Roman pes, are the mostfrequent. This is logical because the pes correspondswith the average length of a Roman foot. Most risersare between 17 and 23 cm high, with the highestmeasuring 30 cm and the peak being 20 cm. This

178

23 Giseke Hopstaken will carry out further research on this typeof threshold.24 For an older type from Augustus’ time in Rome see Rickman1971, 94-5, fig. 21.25 Cf. Rickman 1971, 60-61, fig. 16.26 Adam 1999, 320-321, figs. 730-731.27 Rickman 1971, 68. His statement that this ‘example of thesparing use of travertine in the early Empire when it wasemployed only (sic!) at points of stress’ is rather simplistic. Thecorner columns in the same Horrea for example have an estheticfunction. The same holds good for travertine capitals on tufacolumns. Meiggs 1973, 72.28 Rickman 1971, 48.29 Bakker (ed.) 1999, 150 (ad room 7, west).

179

Fig. 8. Casseggiato dei Molini: threshold (I.iii.1)

Fig. 9. Casseggiato dei Triclini (I.xii.1)

measure is most convenient when a person is goingup and down stairs. The length of the treads, on onestaircase of the same length, are less standardized.They vary from 120 to 180 cm or from 4 to 6 Romanfeet. Most examples vary from 140 to 160 cm, withan average of about 5 Roman feet. This can beexplained by the need for two persons to be able topass each other on the stairs. If the average width ofa person was about 60 cm, then on average about 30cm of space would be free. Usually the blocks ofone or two pieces of travertine are resting on bricksteps. In many cases only the lowest steps or thosewhich would have been visible to the visitor aretravertine, while the remaining steps are brick oreven wood (fig. 9). If the steps are travertine, theirnumber varies from 5 to 23. Usually the staircaserises at an angle of 43-47 degrees.30 In the Baths ofMithras only the superior level of brick layers of theexternal and corresponding stairs carry slabs of trav-ertine (fig. 10). Often times a small vestibule,marked by pivotholes in the threshold, could be

closed off by one or two doors which preceded thereal staircase. If a second staircase rests on a higherlevel or on the first floor, a square landing connectsthe two staircases at a 45 or 90 degree angle. Belowthe staircases were toilets or rooms for storage, asin modern houses. Light and air was afforded by asmall square hole in one of the steps. Large stairs canbe found in the Theater, Schola del Traiano, sometemples and other monumental buildings that saw alot of human traffic. Here the length of the stairsvaries from 400 to 1500 cm. Very small stairs canbe found in subterranean spaces of baths.

4.7 Corner-stones

These small blocks, one or more above each other,often in a symmetrical scheme, are visible on thefrontal corners of doorways (fig. 11) and piers or pil-lars of porticoes (figs. 12 and 13). Sometimes square

180

30 Boersma 1985, 154.

Fig. 10. Bath of Mithras, external staircase (I.xvii.2) Fig. 11. Horrea Epagathiana, corner stones (I.viii.3)

181

Fig. 12. Via della Foce, western side: small corner stones (III.xvi.6)

Fig. 13. Ibidem, small corner stone (detail)

or round holes for the insertion of bars can be foundon the inside of the doorway. The measures of theblocks are not standardized in any of the buildings.They occur most frequently in the Caseggiato delLarario and the Horrea Epagathiana where on aver-age they are 26 to 27 cm long, 15 to 20 cm wide, andc. 14 cm high. But it is interesting to note that the ver-tical distance between the stones is c. one Roman footor a multiple of it. The following table shows thepresence of one or more blocks at the height of oneor more Roman feet. Blocks may be on other levelstoo, at a level different 1/2, 1/3 to 2/3 foot from theindicated heights. If the lowest block is 20 cm, thenthe next block may be 20 cm + one or more feethigh. Not all level-differences can be measured infeet; distances may be 50, 75, 100 and 125 cm.31

Most corner-stones do not have bar holes. Becauseof their irregular, non-standardized form, corner-stones are not needed to strengthen walls. Therefore,they were principally aesthetic and may haveemphasized the status of buildings. They most fre-quently date to the period between c. 65 and 180 A.D.

4.8 Centralized bar stones

Centralized bar stones, dating between c.112 and150 A.D., are visible in a limited number of build-ings: Portico (I.xiv.8; fig. 7), Portico del PiccoloMercato (I.viii.1), Caseggiato dei Misuratori diGrano (I.vii.1) and Horrea Epagathiana (I.viii.3).These buildings with commercial characters are allsituated in Regio I. The bar stones occur on the insidesof doorways or pillars, but never in a symmetricalscheme. The round central holes, often with rests oflead or iron, indicate that bars could be inserted intothis type of stone. The horizontal bar must have beenattached to a vertical one, because there is no corre-sponding bar stone on the opposite side. The levelsof the bar stones are not standardized. They havebeen found at heights of c. 60, 75, 125, 175, 200 and225 cm. Their average length is c. 21 cm and aver-age height c. 15 cm.

4.9 Springers

Horizontal springers or springstones are rare in Ostia.They occur in the Caseggiato del Larario (I.ix.3; c.120 A.D.), in I.xii.4 (fig. 14) and in the entrances ofthe complex of the Case a Giardino (III.ix; c. 128A.D.; fig. 15). Their function may be twofold. First,their sloping sides support a lintel arch of bricks onthe left and right. Second, many examples have apivot hole in the left or right frontal corner corres-ponding exactly with the pivot hole in the thresholdof the same doorway. The level and measures ofspringers are not standardized. In the Casa delLarario (fig. 16) they are c. 300 cm (ten feet) highand in the Case a Giardino they are c. 258 cm high.The average measures are: 58 cm long, 57 cm wide,and c. 42 cm high. The width (c. 2 feet) correspondswith the width of the wall of the doorway. The old-est examples of horizontal springers can be found inthe Forum of Caesar at Rome.

182

31 E.g. in Regio III.xvi.6 (50 and 75 cm); in I.xx.1 (125 cm).

30 60 90 120 150 180 210 240Domus fulminata (III.vii.3) x x xDomus del decumano (III.ii.3) x x xCas. Misuratori (I.vii.1) x xCas. L arario (I.ix.3; fig. 16) x x xCas. balcone ligneo (I.ii.6) x x xCase a Giardino (II.ix.1) x xCasa di Annio (III .xiv.4) x xCas. degli Aurighi (III .x.1) x x xCas. con portico I.xii.7) x xHorrea Epagathiana ( I.viii.3) x xPortico dell’Ercole (IV.ii.2) x xPortico Archi trionf. (V.xi.7) x x

(

Fig. 14. Springers, small corner stones, staircase (I.xii.4)

4.10 Corbels

Corbels occur in at least 13 buildings, mainly betweenc. 100 and c. 170 A.D.32 Corbels carried the woodenjoists, which supported the planks making up thefirst floor. Usually they occur in houses, flats andapartment buildings. They are most frequent in theshops to the right of the Horrea Epagathiana(I.viii.3) and in the Caseggiato del Larario (I.ix.3;fig. 16). Their level is not fixed, but they are on thesame level in the Horrea Epagathiana (288 cmhigh), the Insula del Bacco (I.iv.3) (295 cm high),and in Casa del Bacco (I.vi.2) (379 cm high). Onlythe ends of a corbel are visible because its centralpart is in the wall.33 The extremities are c. 30 cmlong, c. 22 wide, and c. 22 cm high. The distancebetween corbels was not fixed and varies from 105to 285 cm. Within one room the distances are usu-ally identical. Small corbels sustaining marblelatrine benches (usually spolia) are visible in theForica (IV, iv, 4; fig. 17).

4.11 General conclusions and observations

Travertine was used primarily to bear weight, but italso had liminal, marking, corner accentuating, pro-

tective, substituting, layout- and aesthetic and sta-tus-functions. Travertine was first used in Ostia atthe end of the second century B.C. and was mostpopular in the second century A.D., the time of thebiggest building activities. In patrician houses of thethird and fourth centuries A.D., it occurs rarely: onlyin small external stairs, some internal stairs, entrancethresholds, small blocks flanking internal marblethresholds (fig. 18), small isolated blocks with apivot hole (fig. 19), and capital cushions.34 In thesedwellings marble was lavishly used. The contem-poraneous use of marble in bars and baths demon-strates that marble was far more available than inthe first and second centuries A.D. Travertine buil-ding elements were used up until c. 500 A.D., butmainly as spolia. For example corbels were used asfoundation stones, irregular blocks were used as

183

32 In Regio I: i.4; iii.3; iv.3; iv.4; vi.2; viii.3; ix.3. In Regio III:i.8; ix; xiv.4. In Regio IV: ii.2. In Regio V: iii.3.33 Some complete items can be seen in the Baths of Mithras andin IV.vii.1 (used as spolia) and in the southwestern part of theCaseggiato del Larario.34 Becatti 1949, 3, 8, 9, 11-12, 15, 20, 21, 23, 26, 29, 32-33, 35.See also P.F.B. Jongste, Het gebruik van marmer in de Romeinsesamenleving. Diss. Leiden 1995.

Fig. 15. Case a Giardino, entrance, springers (III.ix)

buffer stones on the corners of buildings. The old-est examples of travertine spolia are thresholds anda funerary slab35 reused as foundation stones for thePorta Romana when it was rebuilt c. 100 A.D. (fig. 20).Generally, travertine has more than one function asthe introduction and short survey of architecturalelements has shown above. Before c. 200 A.D. ele-ments like corner-stones may have been status sym-

bols; after c. 200 A.D. travertine lost its prestigiouscharacter, which is confermed by Pensabene’sresearch on capitals.36 Not many standardized build-ing elements have been found. We must concludethat much of it was made ad hoc. However the

184

35 CIL XIV, suppl. I, 5228.36 Cf. Pensabene 1972, 186 (and passim).

Fig. 16. Caseggiato del Larario, springers, small corner stones and corbels (I.ix.3)

following analysis shows that more specifically intombs some standardization was practiced.

5. TRAVERTINE DOORFRAMES IN THE CEMETERIES OFISOLA SACRA AND OF PORTA LAURENTINA AND INOSTIA37

Introduction

Preserved complete or partial travertine doorframesin Ostia are very rare, hence so are specimens oftravertine. Each entrance has its own peculiarities.Travertine doorframes are more numerous in thePorta Laurentina and Isola Sacra cemeteries. Thefirst has six completely preserved doorframes (tomb4, 13, 30, 32, 33, 46) dating from the end of the firstcentury B.C. until c. 100 A.D. Many of the monu-mental tombs of Isola Sacra (c. 100-200 A.D.) havesimilar travertine doorframes.38 This resemblancemakes it worthwhile to investigate whether there isa system: is there any relation between measures,proportions and dating? Was there a preference forspecific measures, even when there is no relationwith dating? Further, did the type of tomb influencethe measures of doorframes? For our research westudied c. forty completely intact monumental door-frames in Isola Sacra. Next it is interesting to com-pare the doorframes of both cemeteries. Togetherthey comprise a period of the first century B.C. untilc. 200 A.D.; the question is whether they had thesame architectural tradition or whether both cemete-ries have to be considered separately. Because of thenumber of doorframes we deal first with the IsolaSacra necropolis hoping to discover a system. Thenthe question rises whether the doorframes of thePorta Laurentina necropolis can be related to it.

Finally, we pay attention to the few, heterogeneousdoorframes in Ostia.

5. 1. The necropolis of Isola Sacra

5. 1.1 General observationsThe necropolis of Isola Sacra lies along the Via Fla-via, a road connecting Portus with Ostia. The dateof this road (c. 70 A.D.) gives a terminus post quemfor the tombs, which apart from some exceptions areturned with their entrance towards the road.39 Theposition of the necropolis is completely in accor-dance with the Roman custom to situate cemeteriesalong or next to a road away from the city.40 Thereare no monumental tombs in isolated position butrows of tombs, separated by paths and open spaces,give the impression of a city, a city of the dead(fig. 21). The buried were members of the middleclass: artisans, shopkeepers, merchants, predomi-nantly liberti and descendants of liberti. Not a sin-gle tomb belonged to members of the upper-class.There is, however, a ‘campus of the poor’, situatedbehind the second row of tombs in the northern zoneof the necropolis, originally with urns, amphoras,fossae in which the unprotected body was buried,wooden chests, terracotta sarcophagi, and tombe allacappuccina.41 Most of these objects are no longer insitu. There are some non-monumental and semi-cylindrical tombs dispersed over the terrain.The necropolis underwent an enormous develop-ment in the second and in the first half of the third

185

37 Studied and written by Natalie Stevens.38 Doorframes of much smaller size, which are present in non-monumental tombs, have been left out of consideration.39 Baldassare 1996, 17. For a map see Pavolini 1983, 260-261.40 Meiggs 1973, 455.41 Baldassare 1996, 23.

Fig. 17. Forica, corbel (IV.iv.4) Fig. 18. Side blocks (V.ii.8)

186

Fig. 19. Isolated block with pivot hole (III.ix.12)

Fig. 20. Porta Romana, substructure of spolia

century A.D. The first rows of tombs were builtdirectly along the Via Flavia (c. 70-100 A.D.). Inthe second phase (Trajanic period) new groups oftombs were placed behind the first rows. In the thirdand fourth phase (Hadrianic-Antonine period, repec-tively third century A.D.) tombs were built againalong the Via Flavia, either using or superimposingon the older rows. Burials ceased at the second halfof the third century A.D.42

Until the Hadrianic period cremation was the burialrite, thereafter inhumation. During some time bothsrites occurred simultaneously. In the Trajan andHadrianic periods the columbarium was the pre-dominant type of family tomb. With the introductionof inhumation, columbaria were adapted by addingspecial niches, arcosoli. In a later phase tombs werebuilt exclusively for inhumation.43

The columbarium type, to which all our completetravertine doorframes belong, can be subdivided ina) tombs with cella (29.3 % ); b) tombs with cellaand with klinai at the outside (14.6 %); c) tombswith cella and enclosure (36.6 %); d) tombs with acombination of b and c (14.6 %) (fig. 22); e) tombswith an upper floor (4.9 %) (fig. 23).44

5.1.2 Travertine doorframesThe c. 40 selected doorframes mainly date from theTrajan and Hadrian-Antonine periods.45 The tombssituated directly alongside the Via Flavia dating backto the third century A.D. have fragmentary doorfra-mes. They are not included in our studies, neither arethe non-accessible tombs to the east of the ViaFlavia.Generally, the selected doorframes consist of the fol-lowing elements (fig. 24): a) two doorposts, flat atthe outside, with an average width of c. 23 cm, adepth of c. 45 cm, and a height of c. 147 cm;46 theleft doorpost (seen from the outside)47 has a barholeat a height of c. 60-100 cm; the space between thetwo doorposts measures 60-90 cm; b) a threshold of

187

42 Baldassare 1996, 17-18.43 Meiggs 1973, 464-465.44 Calza 1940, 63. Category d is our addition.45 The selection consists of doorposts in tomb nrs. 10-11, 13-21,25, 29-31, 42, 54-55, 72, 75-80, 85-87, 89, 92-95, 97.46 The dimensions mentioned are only the most frequent mea-sures.47 Left and right mean: as seen by the onlooker standing in frontof the tomb.

Fig. 21. Isola Sacra (tombs 80-77)

188

Fig. 22. Isola Sacra, tomb 87 Fig. 23. Isola Sacra, tomb 41 (below) and 42 (above)

Fig. 24. Isola Sacra, tomb 11 Fig. 25. Isola Sacra, tomb 89

one block of very varying length, with an u-formedraised edge and a pivot hole in the right frontal cor-ner; and c) a lintel, which, with some exceptions,has the form of a corbel (as frequently used in Ostiacity). In the right frontal corner is a second pivothole, corresponding to the first one in the threshold.The doorposts and the lintel are usually framed bya one-brick layer. Often there is a marble slab witha funerary inscription framed by terracotta, some-times flanked by two slit-windows above the door-frame.

5.1.2.1 DoorpostsThere is no obvious relation between the measuresof the doorposts and the date of the relevant tomb.The smallest width measured is 14 cm, the largest33 cm. During the manufacturing of doorposts theRoman pes (ca. 29.77 cm) has been used, but not ina perfect way.48 Within slight margins each measurecan be converted to Roman measures.49 Of all ourselected doorframes 58.2 % has a width of c. 0.75pes, 24.1 % of c. 0.5 pes and 17.7 % of c. 1.0 pes.Throughout all periods the most frequently chosenwidth is c. 1.5 pedes, that is in 62.5 % of all door-posts. The percentage of widths of 1.75 pedes is

17.25 %. The most frequent height of doorposts isbetween 4.5 and 5 pedes, 28.6 % measuring 4.75pedes and 32.5 % 5 pedes.As for the raised edges, throughout all periods thereseems to have been a preference for a height of 5 cm(41 %; against 24 % of 6 cm and 23 % of 4 cm). Thefirst measure which does have a relation to the date,is the width of the edge. When one compares thepercentages of the various widths occurring in thefirst half of the second century A.D. to the percenta-ges in the second half of the second century A.D.,it can be concluded that the edges become moreslender in the later period. Until 150 A.D. a widthof 15 cm prevails (28.8 % of all doorpost until 150A.D. against 7.7 % after 150 A.D.), after 150 A.D.edges of 13 cm. prevail (30.8 % of all doorpostsafter 150 A.D. against 15.4 % until 150 A.D. ), of11 cm (26.9 % against 9.7 % until 150 A.D.) and of12 cm (23.1 % against 5.8 % until 150 A.D.).

189

48 There is still some uncertainty about the length of the Romanpes. Calculations vary from 29.34 cm (R. de Zwarte, in:BABesch 69 (1994) 129, 142) to 30 cm (Heres 1982, 25). Seealso J.-P. Adam 1989, 41 (29.57 cm). We prefer 29.77 cm (cf.J. de Waele, in: BABesch 59/1 (1984) 1-5).49 We have taken into account quarter and half pedes.

Fig. 26. Ostia, Laurentine cemetery, tomb 32

The measures and the height of the barholes, holesin the left doorpost, making part of the locking sys-tem, are neither bound to a period nor are they stan-dardized. The barholes have a more or less squareform. There is a preference for a width and heightof 7 cm, though 6, 8 and 9 cm also occur frequently.The smallest one measures 4 x 4 cm, the largest 8x 30 cm. The height, measured from the threshold,strongly varies: the lowest point is 56 cm, the highest121 cm. There is preference for a height between 75and 79 cm. Of the 41 barholes 11 (26.8 %) belongto the latter category. To the adjacent categories of70-74 cm and 80-84 cm belong 17.1 % respectively12.2 % (7 respectively 5 barholes); 14.6 % (6 bar-holes) are at the height of 90-94 cm.

5.1.2.2 LintelAs for the lintel, just as in the case of doorposts, norelation can be discovered between measures anddate. The most frequent width is 5 pedes: 34.2 %.Also 5.25 (21 %) and 4.75 pedes (13.2 %) occur.The favourite depth throughout the periods is – justas the depth of doorposts – 1.5 pedes, that is 64 %,while half of this depth is the most frequent height:59 % of the lintels is 0.75 pes high.As for its form the raised edge corresponds with theedge of the threshold (see below) and as measurescorrespond with the edge of the doorposts. Thewidths are between 11 and 19 cm; widths of 13, 14,15 and 16 cm prevail. The height of the edge is inmost cases 5-6 cm. The corbel-like relief protrudesbetween 10 and 24 cm: measures between 16 and17 cm occur most frequently. The narrow edge at thefront of the corbel-like architrave varies in heightfrom 4 to 15 cm. Maybe the edge becomes moreslender during the second half of the second centuryA.D., though there are too few data to be certain.The pivot hole in the right corner, correspondingwith that in the threshold, in which the pivot of thedoor turned, has a diameter of 8-9 cm.

5.1.2.3 ThresholdThe threshold with u-formed edge and a pivot holeof 7-9 cm in diameter has a width between 3.5 and9.75 pedes. Again there is no relation between widthand date. 61.9 % of the measured thresholds has awidth of 4.75-5.25 pedes. Thresholds with a widthof 5.75-7.5 pedes are missing completely. As for thedepth 1.5 and 1.75 pedes (45.5 respectively 27.3 %)have been preferred throughout all periods; the mostfrequent heights are 0.5 pes (44.8 %) and 0.75 pes(41.4 %). The u-shaped edge has a height of 3-7 cm,while 4-6 cm are the most frequently occurring mea-sures. Until 150 A.D. 6 cm was preferred, after 150A.D. 4 cm. The width of the edge varies from 8-18 cm,with a preference for 11, 12, 13, 14 and 15 cm. In the

case of 15 cm 6 of the 7 edges date from before 150A.D.

5.1.2.4 RatiosThe relation between the width of the doorframe andthe width of the facade of the tomb is not fixed. Itvaries from 1:2.9 to 1:4.1. It is not possible to drawconclusions about a development throughout theperiods. In many doorframes the left and right door-post do not have the same width. In 30 % of thedoorframes the width is identical. The right door-jamb is longer in 21.5 % and shorter in 48.5 %. Itis worthwhile to examine how the horizontal ele-ments – threshold and lintel – are related to eachother. The percentage of doorframes having a lintelwith the same length as the threshold is 16.7; 30 %of the lintels is longer and 53.3 % shorter than thethreshold. Although the data are few, it seems thata longer lintel was more in vogue in the first half ofthe second century A.D. (89 % against 11 % in thesecond half of the the second century A.D.). Thewidth of the doorway varies from 1.75 to 3.5 pedes.Most frequently occur the categories 2.5-2.75 (41 %),2.25-2.5 (28 %) and 2.0-2.5 (15 %). A developmentthroughout the periods cannot be discovered. Thewidth of the doorframe, the width of the doorpostsincluded, measures between 3.25 and 4.75 pedes.The studied items are regularly divided over allinterlying quarter pedes, with the exception of 4.0-4.25 and 4.5-4.75 pedes which were preferred in thesecond (31 % against 8 %) respectively the first halfof the second century A.D. (20 % against 0 %). Animportant reason why doorframes differ, is the factthat lintel and threshold sometimes are wider andsometimes not wider than the width of the door-frame. Both types of horizontal elements show gra-dations: they protrude 1) not, 2) a little bit, 3) rathermuch, and 4) much. Horizontal elements both pro-truding over or under the the vertical ones were pre-ferred.

5.1.2.5 Preferred measuresFrom the foregoing it can be deduced that the mostfrequent measures are the following:*doorpost: width: 0.75 pes

depth: 1.5 pedesheight: 5.0 pedes

*threshold: width: 5.0-5.25 pedesdepth: 1.5 pedesheight: 0.75 pes

*lintel: width: 5.0 pedesdepth: 1.5 pedesheight: 0.75 pes

We may conclude that the same preferred measureshave been used for all elements. Thus the question

190

rises whether there are doorframes in Isola Sacrawhich have been built to the standard of the measuresmentioned and whether these standard doorframescan be connected with a tomb type. Within our selec-tion three doorframes completely correspond to thepreferred measures, namely tomb 20, 21 and theenclosure door of tomb 89 (fig. 25). One doorframe(tomb 13) is responding within a few centimetres.It is astonishing that tomb 20-21 and tomb 89 datefrom 160-170 A.D. and tomb 13 from 150-160A.D.50 The number of standard doorframes is toolow to draw hard conclusions. Tomb 13, 20 and 21belong to the columbarium type a, tomb 89 to typec. The type of tomb has not influenced the measuresof the doorframe. Probably the standard measures0.75 x 1.5 x 5.0 (- 5.25) pedes were used during thewhole period of the second until the beginning ofthe third century A.D. with slight variations. Theplacing of prefabricated elements varied per tomb;this gave each facade its own appearance.

5.2. The necropolis of Porta Laurentina

5.2.1 General observationsThe necropolis is situated along the Via Laurentina,which runs from Porta Laurentina in Ostia to thesoutheast. A number of tombs lies along a sidepathand round a square to the east of the Via Laurentina.From c. 50 B.C. to c. 100 A.D., just as in the IsolaSacra necropolis, mainly small merchants and arti-sans, for the most part liberti and some free born,found their last resting place.

5.2.2 Travertine doorframesBecause of the low number of measure data it is notuseful to give percentages like we did for IsolaSacra. We limit ourselves to a comparison betweenthe doorframes of both cemeteries. The tombentrances in the Porta Laurentina cemetery are less

191

50 The dates are proposed by Baldassare 1996.

Fig. 27. Ostia, Laurentine cemetery, tomb 4 Fig. 28. Ostia, Domus delle Gorgoni, doorframe (I.xiii.6)

impressive (fig. 26). No doorposts are higher than4.75 pedes; the measures 3.75 and 4.25 pedes seemto have prevailed. In Isola Sacra the standard heightis larger (5.0 pedes). The widths of threshold andlintel, in Isola Sacra likewise 5.0 pedes, are in somecases 2.75 or 3.5 pedes. These measures are shorterthan the shortest length measured in Isola Sacra. Thecorbel-like lintel which is characteristic for IsolaSacra, occurs only in tomb 4 (50-30 B.C.)51(fig. 27)and tomb 30. The frame of brickstones does notoccur. All the same, the type of doorframe stronglyresembles the type in Isola Sacra: the horizontal ele-ments (threshold and lintel) protrude in most casesand the construction of the doorframes shows greatsimilarities (cf. fig. 27). Most of the pivot holes aresituated at the right side; the barholes are visible inthe left doorpost. Because of the many congruenceswe conclude that there must have been a direct rela-tion between the doorframes in both cemeteries. Thedoorframes in the Porta Laurentina necropolisshould be considered as predecessors of those in theIsola Sacra necropolis.

5.3. Ostia

During our research in Ostia we found two completedoorframes, one monumental porch and some frag-mentary doorframes, with which we shall dealbriefly. Somewhat similar with the types in the IsolaSacra and Porta Laurentina cemeteries – flat door-posts and an u-formed threshold – are 1) a frag-mentary doorframe belonging to the Domus delleGorgoni (I.xiii.6) which may date from the end ofthe third century A.D.;52 2) a doorframe without thelintel in the Caseggiato dei Triclini (I.xii.1) from c.120 A.D.;53 3) the fragmentary entrance of a taberna(V.ii.6), probably from the Severan period.54

Ad 1) The measures of the doorposts (width: 1 pes;depth: 1.75 pedes) and of the threshold width: 5.25pedes; depth: 1.75 pedes) only slightly deviate fromthe average measures in Isola Sacra. Indeed, the

192

51 Pavolini 1983, 238.52 Ibidem, 203.53 Ibidem, 108-109.54 Ibidem, 208.

Fig. 29. Ostia, Tempio dei Fabri navales (III.ii.2) Fig. 30. Cemetery outside Porta Romana, tomb 17

width of the doorframe is larger (6 pedes) and thethreshold has a pivot hole both in the left and theright corner but these differences may be ascribedto different functions. It is remarkable that thethreshold is shorter than the width of the doorframe,a phenomenon which does not occur in the ceme-teries mentioned. Also the setting of the doorpostson the threshold is peculiar; in the frontal part of thethreshold savings have been cut to contain the door-jambs for half (fig. 28).Ad 2) The doorframe in the Caseggiato dei Triclinideviates from those in Isola Sacra insofar it has aconstruction of doorposts in two parts (fig. 9). Theproper doorpost stands on a low sideblock with thesame width (1.5 pedes high). Such side-blocks,albeit without doorposts, are present in the Foricabehind the Caseggiato dei Triclini (3 respectively2.5 pedes high), in the Domus della FortunaAnnonaria (V.ii.8) and in the Insula di Giove eGanimede (I.iv.2). It is not certain whether thesesideblocks really belonged to a doorframe, since theadjacent walls do not show traces of doorposts

which would have been placed against them.Ad 3) The entrance of the taberna has a muchbroader opening (7.5 pedes), which is not surpris-ing in view of its function. Here too sideblocks (2pedes) are present. The width of the remaining door-post is larger than the average doorpost in IsolaSacra or Porta Laurentina. The doorpost, however,has a shorter depth than the sideblock upon whichit is resting.A very curious doorframe is present in the Tempiodei Fabri navales (III.ii.2) from 180-190 A.D., con-sisting of a threshold and lintel, somewhat compa-rable to those in Isola Sacra. Only on places wherebarholes had to be made, blocks of travertine havebeen inserted into the latericium-wall (fig. 29).Of a completely different type is the monumentalentrance to tomb 17 in the necropolis along the ViaOstiense, which probably dates from the secondcentury A.D.55 The gate consists of two doorframes:one exterior and one interior. The external doorframe

193

55 Ibidem, 42. For a map of the cemetery see Pavolini 1983, 38.

Fig. 31. Ibidem, tomb 17, interior doorframe Fig. 32. Ibidem, tomb 17, interior doorframe (detail)

is decorated with continuing reliefbands, in whichthe relief rises towards the outer edges (fig. 30).Parallels for this type can be found in the marbledoorframes of the Terme di Buticoso (I.xiv.8)(122A.D.) and of the Sanctuary of Attis (IV.i.3)(c. 200A.D.) in the Campus Magnae Matris, and in thefragmentary doorframe in the Tempio dell’ AraRotondo (I.xv.6). The marble items, however, do notshow reliefbands rising in outward direction.56 Onthe lintel of tomb 17 is an inscription: H(oc) M(onu-mentum) H(eredes) N(on) [S(equetur)]. Unlike thedoorframes in Isola Sacra the doorposts are placedmore against than into the wall. The raised edge ofthe threshold connects both doorposts. The walls ofthe small vestibule are built in opus latericium. Theceiling is vaulted. The interior doorframe is undec-orated. Its main face is turned to the inside of thetomb (fig. 31). The lintel is a massive block of trav-ertine with an u-like edge which protrudes. The pivothole is – contrary to those in Isola Sacra – in the leftcorner, corresponding with a groove which deepens

toward the wall (fig. 32). This groove served toinsert the door by which the deepest part functionedas a kind of pivot hole.We end with two unique doorframes: the interiormonumental entrance of the Horrea Epagathiana(I.viii.3) (140-150 A.D.)57 and the entrance to a cor-ridor with a dead end at the east side of the Termedel Foro (I.xiii.7). The entrance in the HorreaEpagathiana (fig. 33) is the only doorframe whichcan be called monumental: it is two times biggerthan the monumental doorframes in Isola Sacra. Thedoorposts consist each of two relatively slender ele-ments. The threshold likewise consists of more thanone block. The lintel is in fact a lintel arch with cunei-form blocks. The second entrance mentioned has

194

56 The reliefbands of the external (fragmentary) doorframes inthe Terme del Mitra rise in outward direction. Probably they arenot made of travertine; the material seems to be concrete withrests of shells. These items have been left out of consideration.57 Rickman 1971, 33, fig. 4.

Fig. 33. Ostia, Horrea Epagathiana, doorframe (I.viii.3) Fig. 34. Ostia, Terme del Foro (I.xii.7); travertine door-post

only a doorpost of travertine at the right (fig. 34).This doorpost is incorporated in the east wall of theTerme, consisting of five blocks with a height of 1-2.25 pedes, a width of 1.25 pedes and a depth of2.25 pedes. A superior block is decorated with a pa-tera umbilicata. The doorpost may be reused mate-rial or an element of a (sacral?) building anterior tothe Terme. The left (lost) doorpost has been replacedby a latericium wall; the lintel consists of an archof bipedales and the threshold is missing completely.The dooropening enlarges in upward direction: thesmallest width is 2.75 pedes, the largest 3 pedes.To resume: it may be concluded that the similaritiesbetween the doorframes in Isola Sacra and PortaLaurentina cemeteries and those in Ostia are too fewto assume an identical workshop or tradition.Possible similarities in measures can be explainedby the use of Roman pedes. Similar elements as u-shaped thresholds, pivot holes or flat doorposts aretoo general to assume a direct relation.

5.4 Conclusion

Apart from some exceptions there is no relationbetween measures, proportions and dates of thetravertine doorframes in the necropolis Isola Sacra.Furthermore the type of tomb does not seem to haveinfluenced the measures and proportions. Preferredmeasures, however, have been used which are iden-tical for each of the three elements: 0.75 x 1.5 x 5.0(-5,25) pedes, though deviations from the standardmeasures occur. By variating in the placing of pre-fabicated elements doorframes of each tomb got anindividual proportion. The doorframes in the necrop-olis of Porta Laurentina belong to the same archi-tectural tradition as those of Isola Sacra in spite ofevident differences. In contrast the doorframes inOstia city itself, which are multiform, do not resemblethe doorframes in the two cemeteries mentioned.

BIBLIOGRAPHY

Adam, J.P. 1999, Roman building. Materials andTechniques. London/New York.

Bakker, J.Th. 1994, Living and Working with the Gods.Amsterdam.

Bakker, J.Th. (ed.) 1999, The Mills-Bakeries of Ostia. De-scription and Interpretation. Amsterdam.

Baldassare, I., Bragantini, I., Morselli, C., Taglietti, G.,1996: Necropoli di Porto. Isola Sacra. Roma.

Becatti, G. 1949: Case ostiensi del tardo impero. Roma.Boersma, J.S. 1985, Amoenissima Civitas. Block V.ii at

Ostia: description and analysis of its visible remains.Assen.

Boersma, J.S. 1996, Private Latrines in Ostia: a CaseStudy, BABesch 71, 151-160.

Calza, G., 1940: La necropoli del Porto di Romanell’Isola Sacra. Roma.

Calza, G. Becatti, G., De Angelis d’Ossat, D., Bloch, H.1953, Scavi di Ostia I, Topografia generale. Roma.

Gallina Zevi, A./Claridge, A. 1996, ‘Roman Ostia’ Revisi-ted. London.

Heres, Th.L. 1982, Paries. A Proposal for a DatingSystem of Late-Antique Masonry Structures in Romaand Ostia (A.D. 235-600). Amsterdam.

Hermansen, G. 1981, Ostia, Aspects of Roman City Life.Edmonton.

Mari, Z. 1984, Forma Italiae. Vol. 27. Tibur (pars tertia).Firenze.

Meiggs, R. 19732, Roman Ostia. Oxford.Packer, J.E. 1971, The Insulae of Imperial Ostia (MAAR 31).Pavolini, C. 1983, Ostia. Guide archeologiche Laterza.

Roma/Bari.Pavolini, C. 1986, La vita quotidiana a Ostia. Roma.Pensabene, P. 1973, Scavi di Ostia VII. I capitelli. Roma.Rickman, G. 1971, Roman Granaries and Store

Buildings. Cambridge.SO I = see Calza G. et alii 1953. SO VII = see Pensabene 1973.Saliou, C. 1999, Les trottoirs de Pompéi: une première

approche, BABesch 74, 161-218.Zevi, F. 1976, Monumenti e Aspetti culturali di Ostia

repubblica, in: P. Zanker (ed.), Hellenismus inMittelitalien. Göttingen, 52-83.

FACULTY OF ARCHAEOLOGY,LEIDEN UNIVERSITYPOSTBUS 9515NL-2300 RA LEIDEN

195